Intermittent adherence to Highly Active Antiretroviral Therapy (HAART) allows resistant mutations to emerge, rendering these potentially lifesaving treatments ineffective. In a variety of settings, medication adherence improved when objectively generated feedback around adherence was provided or when providers addressed adherence issues. MEMS-Feedback (MFB) uses patients' MEMS-generated printouts to invite discussion of patients medication-taking. Patients are encouraged to discuss what they thought, how they felt and what they were doing around the times that doses were missed. The therapists' stance, adapted from that in Motivational Enhancement Therapy, involves empathy for difficulties around adherence and support- for patient-generated motivation. When a commitment to improve adherence has been elicited, the therapist works with the patient to problem-solve around barriers to adherence. The common problem of forgetting is addressed by use of individualized reminder cues and a programmable MEMS-cap that beeps at scheduled dose times. The MEMS data is shown to the patient's provider who also conducts a brief version of MFB. This Stage I Behavioral Therapies Development project follows NIDA guidelines. In the first year of this project, MFB will be piloted in 14 patients. Based on this pilot, a manual will be developed that will define the new intervention and guide its implementation. The manual will operationalize and provide examples of therapeutic strategies such as discussion of MEMS printouts, and exploring ambivalence around discussing adherence with providers. Psychotherapy process measures needed to document adherence to the treatment protocol, the clinical competence of therapists, and critical dimensions of the change process will also be developed. Videotaped sessions will be culled to assemble educational materials. These materials will be used in the selection, training, and supervision of a cohort of therapists for a clinical trial. In the clinical trial, 90 patients with MEMS-demonstrated non-adherence during a baseline evaluation will be randomly assigned to 24 weeks of either MFB or a control intervention. The control patients will receive supportive advice around self-reported adherence and advice around the use of cues. The primary outcome measure will be MEMS-measured adherence to correct dose-time. Secondary outcome measures will include cost-effectiveness and viral load.